The antiquity of life on Earth: evidence from sedimentary carbon isotopes

from the 3.52 billion year old Coonterunah Group, Australia.

Roger Buick1*, Jelte P. Harnmeijer1 & David J. Des Marais2

1Dept. Earth & Space Sciences and Astrobiology Program, University of Washington, Seattle

WA 98195-1310, USA

2NASA Ames Research Center, Mailstop 239-4, Moffett Field CA 94035-1000, USA

*Corresponding Author: phone +1-206-543-1013, fax +1-206-543-0489,

buick@ess.washington.edu

1

2

3

4

5

6

7

8

9

10

11

12

13

Abstract

The 3.52 Ga old Coonterunah Group of northwestern Australia contains the oldest

known low-grade metasedimentary rocks. Thus, it can potentially yield the most ancient,

minimally modified, carbon isotopic data pertinent to the early evolution of life. Laminated

carbonates have δ13Ccarb values of -1‰ to -4‰ (mean = -2.9-2.5‰) with δ18OPDB values of -

7‰ to -20‰ (mean = --1176.12‰). Kerogenous cherts have a wide range of δ13Corg values

from -5‰ to -47‰ (mean = --24.024.7‰), but most lie close to a major mode of -25.4-

25.6‰. The mean isotopic difference between bulk reduced (major mode) and oxidized

carbon species (Δ13C) is thus 2212.89‰, very similar toslightly less than that within

individual samples (mean = 23.423.4‰). The slight depletion in δ13Ccarb resembles that in

other Archean sedimentary carbonates associated with banded iron formations and probably

results from precipitation in a stratified ocean where deeper environments derive some

dissolved carbonate from remineralized organic debris. The spread in δ13Corg values probably

partly results from incipient metamorphic resetting, as these rocks were subjected to mid-

greenschist to lowermost-amphibolite facies conditions at which isotopic re-equilibration

starts to occur. After accounting for these effects, Δ13C falls in the range typical of biological

autotrophic fractionation and is dissimilar to any plausible abiotic processes. As these are the

oldest rocks from which a sedimentary Δ13C can be obtained, the results show that

autotrophic, presumably photosynthetic, life became well established in marine environments

during the first billion years of Earth history.

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

Introduction

Despite much scientific effort, it is still unclear how long the Earth has been inhabited

by living organisms. This uncertainty stems from two main problems: the scarcity of suitable

low-grade metasedimentary facies in the early geological record, and the ambiguity of even

the best-preserved relics of early life-forms. The first is partly the fault ofdue to the dearth of

early Archean rocks in general (the older things are, the rarer they become) and partly because

the vagaries of geologic history have subjected most very ancient rocks to considerable

metamorphism and deformation, obliterating or obscuring any biological relics they might

contain. The second results from the apparent simplicity of early organisms, rendering their

physical remains and traces of their activities vulnerable, and difficult to distinguish from

inorganic mimics. Hence, microfossils and stromatolites provide compelling evidence for life

back to around 3.0 Ga (billion years) ago, but before that, all purported physical records of

biological activity have been questioned, by one expert or another, on the grounds of

syngenicity or biogenicity.

The carbon isotope record, based on the preference shown by autotrophic organisms

for the light stable isotope 12C over 13C during carbon fixation, is somewhat more robust, but,

being less prone to metamorphic degradation and younger contamination, as well as being

more readily accessible because data are not so dependent on a lucky find. Even so, the

isotopic record becomes controversial before about 3.45 Ga, the age of the oldest very low-

grade sedimentary rocks in Australia’s the Warrawoona Group of Australia and the South

Africa’s Onverwacht Group of southern Africa. Specifically, debate has focussed on the age

and origin of carbonaceous matter in the Isua and Akilia areas of Greenland. These upper-

amphibolite to granulite facies rocks are older than 3.7 Ga and contain massive carbonates and

graphitic gneisses that show carbon isotopic fractionations not inconsistent with a biological

origin followed by later metamorphic resetting (Schidlowski, 1988). However, the carbonates

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

have recently been reinterpreted as metasomatic derivatives of ultramafic igneous rocks, not

sediments (Rose et al., 1996), and hence their isotopic values reflect abiotic processes only.

Graphite yields a wide range of δ13Corg values, from as heavy as -3‰ in ankerite-quartz-

cummingtonite schist (Ueno et al., 2002) to as light as -45‰ from minute inclusions in apatite

crystals (Mojzsis et al., 1996). To explain this spread, it has been argued that the heavier

values represent metamorphically reset biological ratios, whereas the lighter values are closer

to an original autotrophic signature (Schidlowski et al., 1979; Ueno et al., 2002). But it is

now evident that much of the graphite at Isua was produced abiotically by metamorphic

reduction of metasomatic siderite (van Zuilen et al., 2002) and doubtful whether any graphite

at all occurs at Akilia (Lepland et al., 2005; Moorbath, 2005). The most likely Isua graphite to

be both biological and ancient (~3.8 Ga) is moderately abundant (TOC = 0.1-0.9 wt.%),

resides in turbiditic metasediments that are not associated with metasomatic carbonate and

ranges in δ13Corg from –20 to -14‰ (Rosing, 1999). However, meteoritic organic carbon from

carbonaceous chondrites and tarry residues from prebiotic synthesis experiments (Miller-

Urey, Fischer-Tropsch) show a similar range of values. Moreover, in the absence of a reliable

sedimentary δ13Ccarb record with which to compare these Isua organic carbon values,makes it

is impossible to prove beyond all reasonable doubt that the total isotopic fractionation reflects

biological autotrophy (Buick, 2001).

Here we present δ13C analyses of organic and carbonate carbon from a recently

discovered succession of slightly younger and better preserved sedimentary rocks from the

Pilbara Craton in Australia. These yield much less equivocal data which extend the

incontrovertible record of life on Earth back to the first billion years of this planet's history.

Geological Setting

The rocks examined here belong to the Coonterunah Group, the basal stratigraphic

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

unit in Archean Pilbara Craton of northwestern Australia (Fig. 1). The ~5.5 km thick

succession, about 5.5 km thick as preserved but truncated at the base by slightly younger

granitoid intrusions, is dominated by tholeiitic basalts with subordinate felsic volcanics and

volcanogenic sediments, magnesian and komatiitic basalts, and synvolcanic gabbroic

intrusions (Buick et al., 1995; Green et al., 2000). From the widespread presence of pillows,

sparse amygdales in basalts, rare breccias in dacites and paucity of clastic sediments, the

depositional environment was apparently deep marine. Dacites and rhyolites from near the

top of the succession have been dated by SHRIMP U-Pb in zircon techniques as 3515-3517

Ma old (Buick et al., 1995). The metamorphic grade varies along the exposed strike length of

~75 km from mid-greenschist facies to lower amphibolite facies (Green et al., 2000), with a

tectonic fabric only developed in the higher-grade areas. In most areas, deformation has

consisted of nothing more than early open folding, followed by later block faulting and

subsequent then tilting to the present subvertical bedding planes dips. Following the erosional

unconformity and After erosion creating an upper unconformity surface and thensubsequent

deposition of the very low-grade and minimally deformed Warrawoona Group at ~3.4 Ga,

very little has disturbed their repose (Buick et al., 1995). Thus, they are the oldest known

low-grade rocks on Earth and as such, are an excellent place to search for isotopic records of

early life.

Ferruginous chert beds (Fig. 21a) form thin (~0.3m thick) layers between basalt flows

and thick (10-30m) units between different volcanic lithologies. They have white and red or

black banding on a centimeter scale and a poorly defined lamination on a millimeter scale.

The layering is planar to undulating with diffuse boundaries, showing no evidence of clastic

sedimentary structures or textures. However, the chert is clearly sedimentary and not

secondary as the laminae drape topographic highs on underlying basalt flows, thickening into

troughs and thinning over ridges (Fig. 21b). At a microscopic scale, all layers are

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

predominantly composed of equigranular microquartz, with the colored bands and laminae

containing varying amounts (up to 25%) of very fine-grained (~0.1 mm) hematite or

magnetite. Laterally gradational contacts indicate that the hematitic layers are secondary

modifications of originally magnetitic bands. Trace amounts of organic carbon exist in the

magnetitic layers, but TOC is generally too low for successful extraction by acid digestion

without risk of modern contamination. No detrital clastic grains have been observed in thin-

sections. Thus, the cherts are best interpreted as interflow sediments resembling banded iron-

formation that were deposited during times of volcanic quiescence under very low energy

conditions with miniscule levels of clastic sedimentation. Their intimate association with

surrounding volcanic rocks indicates that they formed in the same environmental setting; i.e.

in deep marine conditions.

Associated with some of the chert beds are thin (5-30 cm) beds of massive to plane-

laminated (2-5 mm) carbonate. In most places, these have been secondarily silicified during

Tertiary weathering, resulting in so they crop out as lenseses and podss within layers of

amorphous brown silcrete, with interfingering contacts and grain overgrowths showing that

the carbonate was indeed the primary phase. The carbonate is now quite coarse grained (0.5-1

mm) and composed of equant grey-brown crystals, but this is almost certainly the result of

metamorphic recrystallization and it was presumably finer originally. Along some contacts

with adjacent siliceous beds, carbonate is interlayered with quartz-grunerite-magnetite

laminae, gradually passing into quartz-magnetite chert. The present mineralogy of the

carbonate layers is ferroan dolomite and calcite, identified bythrough stainingmicroprobe

analysis. As in the cherts, no evidence of clastic sedimentary structures or textures has been

observed, and on the microscopic scale, no detrital mineral grains have been found in thin-

sections of the carbonate rocks. So, the depositional environment of the carbonates is best

interpreted as similar to the interbedded cherts and volcanics; i.e. low energy chemical

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

sediments formed in a deep-water submarine setting. Unlike the older Isua carbonates from

Greenland, there are no cross-cutting carbonate veins and adjacent mafic igneous igneous

rocks show only very limited carbonate alteration.

Methods

Surface samples were collected from minimally metamorphosed parts of the

Coonterunah Group, as far as possible from igneous intrusions and cross-cutting faults.

Exterior weathering was removed with a diamond saw and the rocks were then coarsely

crushed. Fragments were etched in HCl and HF to remove any surface contamination and

then finely crushed, in a steel puck mill for carbonates and by agate pestle and mortar for

cherts. Kerogen was isolated from powders with higher TOC by demineralizing with HF and

then removing secondary fluorides with H3BO3, a modification of the technique of Robl and

Davis (1993). Some c Carbonates were reacted under vacuum with HPO4 in sealed quartz-

glass Y-tubes at 100°C for 24 hours, while others were reacted and analysed at 80 °C using a

Kiel-III carbonate mass-spectrometer. Cherts were decarbonated with warm concentrated

HCl. Some cherts and kerogen isolates were and then combusted under vacuum in sealed

quartz-glass bombs with CuO at 800°C for 4 hours. Kerogen isolates were similarly

combusted. In, with the all cases, the resulting CO2 was purified cryogenically in a vacuum

gas-distillation line and then isotopically analysed on a Nuclide isotope-ratio mass-

spectrometer. Others were analysed in a Thermo Finnigan MAT 253. Results are expressed

in δ13C or δ18O notation as per mille (‰) values relative to the PDB standard. For oxygen

isotopes in carbonates, values were calculated using a published calcite fractionation factor at

the appropriate temperature.

Resultsfor 100°C.

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

Results

δ13Corg values are tabulated in Table 1. It is evident that there is a major mode centered

about -25‰ and, with a minor group around -40‰ and a spread skewed towardsof heavier

values ranging from -2015‰ to -5‰. The tight clustering of values around the major mode

(-25.4-25.4‰, corrected for replicates) indicates that it is of primary significance, because

low-grade metamorphism has inconsistent isotopic resetting effects depending on local

conditions and tends to smear the distribution between original sedimentary and bulk earth

values. Moreover, for high TOC samples where a δ13Corg value could be obtained from

isolated kerogen, all fall within the major modal population, suggesting that metamorphic

resetting has only influenced kerogen-poor samples. Thus, the most reasonable interpretation

is that the major mode represents mildly reset original values and therefore is an approximate

indicator of the bulk δ13Corg deposited during sedimentation. If so, it is roughly equivalent to

the values observed in modern sediments where carbon isotopic fractionation is controlled by

RuBisCO-modulated carbon-fixation. Though the spread in values around this particular

mode gets as light as -31‰, it is unlikely that the original values were very much lighter than

this, as maturational hydrocarbon expulsion probably did not shift values by much more than

a couple of parts per mille, even given the extremely low TOC percentages (Buick et al.,

1998). The minor light mode may highlight microbial remineralization of organic matter

during diagenesis and the heavy outliers may represent metamorphic resetting (see

Discussion).

Carbonates show a much narrower range of isotopic values than kerogen. δ13Ccarb

ranges between -1‰ and -4‰, with a mean of 2.-2.95‰ (Table 2). Values are strongly

skewed towards lighter valuesevenly spread about this mean with no obvious subpopulations,

unlike the δ13Corg distribution. . This tighter distribution probably reflects the much higher

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

amounts of carbonate in the samples (CO2 yields average 42.1 µmoles) compared with

organic carbon (CO2 yields average 3.5µmoles). In such circumstances, metamorphic

resetting is likely to have been minimal. Diagenetic alteration may have been significant, But

because the carbonate beds are thin and volumetrically minor within the entire stratigraphic

succession., diagenetic alteration may have been significant. Interactions between

sedimentary carbonates and diagenetic fluids can be monitored in very low-grade terrains by

δ18O values. In the Coonterunah carbonates, these range from -7‰ to -20‰ about a mean of -

17.1-16.2‰. Though it is unknown exactly what marine Archean δ18O values were, it seems

likely that they became lighter back through time (Veizer et al., 1989). Diagenetic alteration

in most sedimentary environments, excluding evaporitic basins, resets the marine signature to

even lighter values. This would suggest that the closest Coonterunah δ18O value to

contemporary seawater was around -7‰ and that the lighter values represent diagenetic fluid-

rock interaction. If so, then original δ13Ccarb values were probably not altered much during

diagenesis, as lighter δ18O values are not correlated with either extremity of the δ13Ccarb

distribution (Fig. 3, inset4). So, while diagenetic and metamorphic fluid-rock interactions

may have occurred, they do not appear to have substantially modified δ13Ccarb.

Δ13C (, the difference between δ13Ccarb -and δ13Corg,) is an approximate measure of the

fractionation between oxidized and reduced carbon species. Given the minor uncertainty

about exact original isotopic compositions because of the effects of diagenesis and

metamorphism, this is a more appropriate measure to use than the precise fractionation figure

εTOC. As δ13Ccarb values form a single homogeneous population, their mean of -2.-2.95‰ can

be used in the calculation. For δ13Corg, the corrected mean of values clustered around the

major mode (-25.4‰-25.6) is used, as this most closely represents the original material

deposited. Hence, bulk Δ13C is 22.722.9‰, very similar to the modern marine value of

~22‰. Though diagenesis and metamorphism have probably reduced the Coonterunah Δ13C

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

value, this was probably only by a few per mille. This is shown in the few kerogenous

carbonates in which Δ13C can be determined for individual samples (Table 3), where any re-

equilibration between reduced and oxidized carbon species should be most pronounced.

Collectively, these indicate a mean Δ13C of 23.423.4‰, closely comparable to the bulk Δ13C

value. This confirms the validity of the latter value and demonstrates that, for most samples,

diagenetic and metamorphic isotopic homogenization has been limited in most samples.

Discussion

An integral problem in any paleobiological study of particularly ancient rocks is

assessing whether the feature under examination, be it isotopes or organisms, is indigenous

and syngenetic. Any rock with an extremely long geologic history has had ample opportunity

for post-depositional contamination, and where contents of the material in question are

extremely low, the chances of modern adulteration are high. Both of these problems have

been encountered in previous organic and isotopic studies of the ~3.8 Ga gniesses of

Greenland, with modern organic contaminants, once thought to be syngenetic, occurring

abundantly in surface rocks, and massive carbonates, formerly believed to be sedimentary,

forming metasomatically near ultramafic igneous rocks. In the case of the Coonterunah rocks,

however, such difficulties were not encountered. The Coonterunah carbonates show

sedimentary lamination, are interbedded with siliceous sedimentary facies and are kilometres

removed from any source of ultramafic metasomatism. The kerogen in the Coonterunah

cherts is clearly indigenous because it can be extracted by acid digestion, yielding similar

isotopic values to in situ treatments that have been carefully manipulated to exclude the

possibility of surface contamination. Moreover, carbonate samples also contain kerogen

clearly embedded in metamorphosed ferroan dolomite and calcite crystals (Fig. 23c),

indicating its premetamorphic emplacement.

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

The spread of δ13Corg values provides further evidence for the ancient origin of the

Coonterunah organic matter, because if it were modern superficial contamination or ancient

post-metamorphic adulteration, it should have a homogeneous isotopic composition. Instead,

the isotopic diversity is best interpreted as representing either metamorphic resetting or

metabolic disparity. As δ13Corg ranges between -47‰ and -5‰ with a mode around -25‰, a

purely biological explanation would have to invoke three distinct mechanisms of isotopic

fractionation. The -25‰ mode could be the product of oxygenic photosynthesis by

cyanobacteria using RuBisCO which imparts a ~25‰ total fractionation from dissolved CO2,

as seen here. Some of this primary production could then have been remineralized by

methanogenic Archaea, which have a cellular biomass markedly lighter than the consumed

organic matter, yielding the few values between -37‰ and -47‰. The heavy values between

-15‰ and -5‰ could be the result of anoxygenic photosynthesis by green or purple sulphur

bacteria, an autotrophic process with a smaller associated isotopic fractionation than its

oxygenic equivalent. However, the range and extremity of heavy values observed make this

rather unlikely as no form of photosynthesis yields organic matter heavier than dissolved CO2,

which would have been around -10‰ for carbonates near 0‰ (Des Marais, 1997). Instead, it

is more plausible that differing degrees of maturational decarbonation of light hydrocarbons

and metamorphic re-equilibration with heavy carbonates were responsible for the spread of

heavy values. Though attempts were made to sample only lower grade rocks, the regional

metamorphic grade, mid-greenschist to lower amphibolite facies, is right at the point where

profound isotopic adjustments start to occur (Hayes et al., 1983; Des Marais et al., 1992; Des

Marais, 1997). It is unlikely that the population around the mode of -25‰ also represents

metamorphically reset kerogen that was originally very much lighter, of similar isotopic

composition to the light subpopulation near -40‰, as the major modal cluster is large and

distinct, without a skewed tail trailing towards the lighter mode.

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

The carbonates show a distinct similarity to younger Archean carbonates associated

with BIFs. A trend in δ13Ccarb values from close to 0‰ to lighter values to -4‰ is evident in

the Hamersley Group (Kaufman et al., 1990), where it has been attributed to deposition in

deeper water below a redox chemocline with increasing proportions of bicarbonate derived

from recycled light organic matter. As the Coonterunah carbonates were apparently deposited

in a similar environment and are associated with magnetite-grunerite cherts, a similar origin

could be invoked. A metamorphic origin for the trend is unlikely, as the carbonates contain

little kerogen and have many orders of magnitude more of the oxidized species than the

reduced, providing a robust buffer to isotopic re-equilibration. If so, the initial oceanic δ13Ccarb

ratio was probably around the heaviest δ13Ccarb value observed, about -1‰. This would accord

with estimates of near 0‰ obtained from somewhat younger Archaean shallow-marine

carbonates (Veizer et al., 1989).

It is possible to erect a set of criteria for recognizing biogenic signatures in carbon

isotopes very early in Earth's history when abiotic contributions may have been more

significant than now. These are listed in order of increasing rigour, with only criterion 5,

where both reduced and oxidized species are congruently preserved together, representing the

highest likelihood of biogenicity. As most of the criteria are based on statistical parameters,

sample numbers (n > 10) and sizes (CO2 yield > 0.3 µmoles) have to be sufficiently large for

robust analysis.

1. Organic negativity - because carbonaceous chondritic meteorites, mantle-derived diamonds

and prebiotic organic syntheses all produce solid organic carbon with δ13Corg values in

the range -15‰ to 0‰, values should be lighter than this range (<-15‰), especially at

low TOC (<0.1%);

2. Kerogen consistency - because abiogenic organic carbon tends to be broadly spread about a

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

mean, the standard deviation should be low (<5) indicating high kurtosis;

3. Carbonate coherence - because biogenic fractionations over the past 800 Ma have ranged

between 22‰ and 32‰ (Hayes et al., 1999) but inorganic carbonates and organics are

generally fractionated less than this, the gross bulk fractionation between δ13Corg and

δ13Ccarb (Δ13C) from a single suite of rocks should be >20‰;

4. Complementary skewness - because metamorphic re-equilibration tends to skew the

isotopic distribution to one side of the mean, if any skewness occurs, it should be

positive around δ13Corg modes and negative for δ13Ccarb;

5. Fractionation congruency - because metamorphic resetting should be most evident in rocks

with coexisting kerogen and carbonate, the standard deviation for the population of

individual sample Δ13C values should be low (<5).

The Coonterunah isotopic values satisfy all 5 of these criteria, so they evidently indicate

biogenic carbon isotopic fractionation.

It is thus reasonable to consider the sorts of metabolic processes that might have been

operative in early Archean ecosystems. The Δ13C fractionation clearly represents autotrophic

carbon-fixation which, given its magnitude and moderate metamorphic resetting, probably

signifies the Calvin-Benson cycle using the enzyme ribulose-bisphosphate carboxylase-

oxygenase (RuBisCO). This, however, need not have been oxygenic photosynthesis; indeed,

it need not have been photosynthesis at all. Many forms of autotrophy employ this

mechanism for carbon fixation, and so other geochemical evidence (e.g. Hayes, 1983; Buick,

1992; Brocks et al., 1999) is needed before oxygenic photosynthesis can be recognized. The

minor modes in the δ13Corg distribution may also represent distinct metabolic processes. The

light mode possibly indicates methanogenesis, as values lighter than -40‰ are beyond the

range of fractionation imparted by any form of photosynthesis but are characteristic of

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

biogenic methane production. Given the sparse data, it is impossible to determine whether

this methanogenesis relied on the heterotrophic breakdown of primary autotrophic production

(organic recycling) or autotrophic redox reactions involving simple gas species

(chemolithotrophy). It does not necessarily also imply oxidative methylotrophy as a means of

fixing the light methane isotopic signature into biomass, because one class of methanogens,

the methylotrophic methanogens which split methyl groups from simple methylated

molecules to produce methane (Summons et al., 1999). The heavy mode, as mentioned

above, probably does not reflect the activities of green or purple sulfur bacteria metabolizing

photosynthetically, which tend to produce organic matter lighter than -10‰. Given the broad

spread of values and their extreme enrichment in 13C, this mode is more likely to represent

metamorphosed organic matter initially derived from other sources.

The secular distribution of sedimentary carbon isotope values for rocks older than 1

Ga3 billion years old is plotted in Fig. 45. It can be seen that the Coonterunah kerogens and

carbonates are not too dissimilar to those from other early Archaean successions. The

carbonates are a bit lighter than those from shallow marine facies, such as the ~3.46 Ga

Warrawoona Group, but are similar to those derived from deeper-water BIF successions. This

difference is also manifest in mineralogy, with the shallow marine carbonates dominatly

composed of dolomite, whereas the deeper water facies, including the Coonterunah

carbonates, are calcitic. The Coonterunah kerogens are a bit heavier than most comparable

early Archaean organic carbon, though not markedly so. This may reflect their higher

metamorphic grade than most, but as the coexisting kerogen-carbonate pairs show little sign

of metamorphic re-equilibration, this may instead be a primary environmental or biological

indicator. As there are several other deep-water successions represented in the early Archaean

carbon isotope record, it seems unlikely to be a simple factor of depositional environment. It

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

could possibly indicate lower levels of dissolved CO2 in the water mass, as isotopic

discrimination during microbial autotrophy diminishes with CO2 availability (e.g. Calder &

Parker, 1978; Des Marais et al., 1989). It might also indicate a greater input of biomass from

non-cyanobacterial microbial autotrophs (e.g. Ruby et al., 1987; Des Marais et al., 1989).

Lastly, it may simply reflect lower levels of methanogenic recycling than in other Archaean

successions (e.g. Hayes, 1983; 1994), though the presence of an apparently methanogenic

light mode in the Coonterunah kerogen data would suggest otherwise. Regardless, a rather

more refined instrument than carbon isotopes is needed to resolve the problem.

When comparing the Coonterunah with the ~3.8 Ga Itsaq data, it is striking that

organic polymodality is a common feature. If the Itsaq data does indeed, in whole or in part,

represent biological activity, then this perhaps reveals that early Archaean microbial

ecosystems had more complex isotopic partitioning than occurs in younger settings. At

present, carbon isotopic signatures are overwhelmingly dominated by biomass generated by

oxygenic photosynthesis, producing unimodal isotopic signatures. As the secondary modes in

Archaean isotopic distributions possibly represent diverse autotrophic and heterotrophic

pathways, this suggests that biomass generation and recycling was not so simple. As many of

these possible pathways are anaerobic or microaerophilic, this scenario would accord with the

model of a less oxygenated early Earth that is widely, but not universally, upheld.

At present, it seems unlikely that older incontrovertible carbon isotope evidence for

the existence of primordial life will be found in rocks that are currently known. All more

ancient successions lack coexisting sedimentary carbonates and kerogen, though they may

have one or the other. Moreover, they have all suffered greater metamorphism and

deformation, beyond the degree where substantial resetting of carbon isotopic values applies.

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

Morphological evidence for life is very unlikely to survive such conditions. Though it is not

inconceivable that older low-grade rocks will be discovered; indeed the Coonterunah Group

was not recognized as being older and better preserved than other Pilbara greenstones until

1995; the co-occurrence of sedimentary carbonates and kerogens is not overly common in

Archean successions generally. Though many Archean terrains are still poorly mapped and

dated, it is likely that particularly ancient low-grade successions remain to be discovered only

in cratons that are generally mildly metamorphosed and very old. Thus, this Coonterunah

data might be as old as it gets. If so, it would be unfortunate for science, because most of the

really interesting questions about the early evolution of life lie in the preceding billion years

of Earth's history, when life itself began, diversified and developed the complex metabolic

pathways already in evidence by the dawn of the biogeochemical record.

359

360

361

362

363

364

365

366

367

368

369

370

371

Acknowledgements

We thank Anne Tharpe for mass-spectrometry, Ian O'Brien and Owen Thomas for

technical assistance, and Michael Green for geological information. This study was supported

by Australian Research Council Large Grants, NASA Exobiology grant EXB03-0000-

0017 and the NASA Astrobiology Institute (RB).

372

373

374

375

376

377

References

Brocks, J.J., Logan, G.R., Buick, R. & Summons, R.E. (1999) Science 285, 1033-1036.

Buick, R. (1992) Science 255, 74-77.

Buick, R. (2001) in Palaeobiology II, eds. Briggs, D.E.G. & Crowther, P.R.

(Blackwell, Oxford), pp. 13-21.

Buick, R., Des Marais, D.J. & Knoll, A.H. (1995) Chem. Geol. 123, 153-171.

Buick, R., Rasmussen, B. & Krapez, B. (1998) AAPG Bull. 82, 50-69.

Buick, R., Thornett, J.R., McNaughton, N.J., Smith, J., Barley, M.E. & Savage, M. (1995)

Nature 375, 574-577.

Calder, J.A. & Parker, P.L. (1978) Geochim. Cosmochim. Acta 37, 133-140.

Des Marais, D.J. (1997) Org. Geochem. 27, 185-193.

Des Marais, D.J., Cohen, Y., Nguyen, H., Cheatham, M., Cheatham, T. & Munoz, E. (1989)

in Microbial Mats: Physiological Ecology of Benthic Microbial Communities, eds.

Cohen, Y. & Rosenberg, E. (American Society for Microbiology, Washington), pp.

191-205.

Des Marais, D.J., Strauss, H., Summons, R.E. & Hayes, J.M. (1992) Nature 359, 605-609.

Eiler, J.M., Mojzsis, S.J. & Arrhenius, G. (1997) Nature 386, 665.

Green, M.J., Sylvester, P.J. & Buick, R. (2000) Tectonophys 322, 69-88.

Hayes, J.M. (1983) in Earth's Earliest Biosphere, ed. Schopf, J.W. (Princeton Univ. Press,

Princeton), pp. 291-301.

Hayes, J.M. (1994) in Early Life on Earth Nobel Symposium, ed. Bengtson, S. (Columbia

Univ. Press, New York), pp. 220-236.

Hayes, J.M., Kaplan, I.R. & Wedeking, K. (1983) in Earth's Earliest Biosphere, ed. Schopf,

J.W. (Princeton Univ. Press, Princeton), pp. 93-134.

Hayes, J.M., Strauss, H. & Kaufman, A.J. (1999) Chem. Geol. 161, 103-125.

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

Kaufman, A.J., Hayes, J.M. & Klein, C. (1990) Geochim. Cosmochim. Acta

54, 3461-3473.

Lepland, A., van Zuilen, M.A., Arrhenius, G., Whitehouse, M.J. & Fedo, C.M.

(2005) Geology 33, 77-79.

Lowe, D.R. (1994) Geology 22, 387-390.

Mojzsis, S.J., Arrhenius, G., McKeegan, K.D., Harrison, T.M., Nutman, A.P. & Friend,

C.R.L. (1996) Nature 384, 55-59.

Moorbath, S. (2005) Nature 434, 155.

Robl, T.L. & Davis, B.H. (1993) Org. Geochem. 20, 249-255.

Rose, N.M., Rosing, M.T. & Bridgwater, D. (1996) Am. J. Sci. 296, 1004-1044.

Rosing, M.T., Rose, N.M., Bridgwater, D. & Thomsen, H.S. (1996) Geology 24, 43-46.

Rosing, M.T. (1999) Science 283, 674-676.

Ruby, E.G., Jannasch, H.W. & Deuser, W.G. (1987) App. Microbiol. 53, 1940-1943.

Schidlowski, M. (1988) Nature 333, 313-318.

Schidlowski, M., Appel, P.W.U., Eichmann, R. & Junge, E. (1979) Geochim. Cosmochim.

Acta 43, 189-199.

Summons, R.E., Franzmann, P.D. & Nichols P.D. (1998) Org. Geochem. 28,

465-475.

Ueno, Y., Yurimoto, H., Yoshioka, H., Komiya, T. & Maruyama, S. (2002)

Geochim. Cosmochim. Acta 66, 1257-1268.

van Zuilen, M.A., Lepland, A. & Arrhenius G. (2002) Nature 418, 627-630.

Veizer, J., Hoefs, J., Lowe, D.R. & Thurston P.C. (1989) Geochim. Cosmochim. Acta 53, 859-

871.

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

Figures

Figure 1. Geological map of the Coonterunah Group, showing sample sites.

Figure 2. Fig. 2. a) Laminated carbonate in outcrop.Banded chert in outcrop, b) Banded chert

laminae draping surface irregularities in the underlying basalt flow, bc) Banded chert

in outcrop,, showing interstitial pelite laminae, d) Laminated carbonate in outcrop, e)

Basaltic pillow breccia with interstitial laminated carbonatec) Thin-section of

laminated carbonate. d) Thin-section of chert-hosted kerogena) Thin-section of banded

chert, b) Thin-section of laminated carbonate, c) Kerogen within rhombic

metamorphosed calcite crystal.

.

Figure 3. Carbon isotopic data from Coonterunah Group. Inset shows cross-plot of δ18Ocarb vs.

δ13Ccarb.

Figure 4. Secular trends in carbon isotopes before 1 Ga.

Supplementary Tables

Table 1. Organic carbon isotopic data: suffix “K” in sample # indicates macerated kerogen,

others are decarbonated whole-rock analyses, suffix “rep” indicates replicate analysis,

suffix “lam” indicated microdremel laminar analysis. Laminar analyses are treated as

distinct samples in calculation of averages.

Table 2. Carbonate isotopic data.

Table 3. Δ13C data.

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

Table 1. Coonterunah kerogen δ13Corg data.Rock-type

Sample Size (mg)

CO2

yield (µmoles)

δ13Corg

(‰, VPDB)

± σ (‰) Sample average

(‰)

Che

rt-h

oste

d ke

roge

n

70615 249.6 2.9 -8.13 0.05-7.2570615 rep1 196.3 1.11 -6.36 0.11

70615K 1.2 0.38 -22.76 0.11 -22.7670648 207 0.57 -27.52 0.11 -27.5270648K 0.4 0.58 -27.66 0.12 -27.6693007 200.5 0.36 -27.91 0.11 -27.9194052K 1.4 31.99 -28.87 0.08 -28.8794053 5sec 200.1 0.60 -30.50 0.09

-30.56

94053 10sec 195.4 0.72 -31.11 0.0794053 15sec 221.9 1.37 -31.86 0.0894053 20sec 220.3 1.23 -29.06 0.0894053 25sec 199.9 0.88 -30.29 0.0796014 205.4 1.47 -22.8 0.06 -22.8096016 218.6 2.2 -14.48 0.11 -14.4896017B 208.3 14.87 -4.72 0.08 -4.7296018 197.3 1.27 -27.3 0.08 -27.3096019A 99.2 0.81 -12.55 0.08

-11.5396019A rep1 196.7 1.47 -10.5 0.0796020C 105.7 0.49 -22.09 0.08

-20.9396020C rep1 201.7 0.82 -19.76 0.0796022C 213.5 2.7 -10.92 0.08 -10.9296031 206 0.61 -25.42 0.1196031 rep1 202.6 0.58 -24.31 0.16 -24.8797020A* 225.3 7.13 -46.66 0.08 -46.6697020B* 200.0 4.93 -41.07 0.04

-39.2997020B rep1 220.6 3.73 -37.50 0.0797021* 204.5 9.78 -19.5 0.09 -19.50PC03-104 0.8 20.70 -23.78 0.38 -23.78PC03-119 2.8 0.20 -23.96 0.13

-23.59PC03-119 rep1 9.9 0.31 -23.22 0.13PC04-005 13.1 1.77 -25.07 0.37

-25.27PC04-005 rep1 6.1 1.06 -26.01 0.37PC04-005 rep2 10.0 0.94 -24.75 0.37PC04-113 6.3 2.59 -17.95 0.38

-17.16PC04-113 rep1 19.4 6.95 -16.61 0.38PC04-113 rep2 14.7 5.32 -16.91 0.38Mean -23.33 9.21 -22.97

Car

bona

te-

host

ed

kero

gen 93006K 9.9 0.98 -26.65 0.08

-24.8893006K rep1 11.4 1.12 -23.1 0.0496045AK 17.2 2.19 -24.27 0.09 -23.5796045AK rep1 12.4 4.72 -22.87 0.07

449

96045BK 4.2 7.24 -28.77 0.06 -28.77PC05-020A blk 5.0 3.90 -28.77 0.12 -28.77PC05-020B lam1 27.8 5.10 -25.34 0.12 -25.34PC05-020B lam2 24.1 5.17 -27.34 0.12 -27.34PC05-020B lam3 24.6 5.76 -26.49 0.65 -26.49PC05-020B lam4 15.2 4.22 -24.29 0.35 -24.29PC05-020B lam5 15.2 15.53 -29.63 0.35 -29.63PC05-020C lam1 32.0 5.57 -27.21 0.12 -27.21PC05-020C lam2 14.4 1.93 -27.67 0.12 -27.67PC05-020C lam3 21.0 2.08 -27.03 0.65 -27.03PC05-020C lam4 23.2 4.83 -26.26 0.35 -26.26PC05-020C lam5 22.5 4.37 -26.33 0.35 -26.33PC05-020C lam6 24.4 5.46 -27.11 0.35 -27.11PC05-020C lam7 8.0 0.68 -24.53 0.09 -24.53PC05-020C lam8 10.0 0.76 -23.24 0.09 -23.24PC05-021 lam1 5.7 2.29 -25.17 0.12 -25.17PC05-021 lam2 11.7 2.12 -31.13 0.12 -31.13PC05-021 lam3 5.5 2.28 -24.08 0.65 -24.08PC05-021 lam4 2.9 1.54 -24.20 0.35 -24.20PC05-021 lam5 0.6 0.31 -22.64 0.09 -22.64PC05-021 lam6 1.0 0.29 -21.32 0.09 -21.32PC06-031 lam1 8.3 0.62 -19.22 0.18 -19.22PC06-031 lam2 1.6 0.35 -23.28 0.18 -23.28PC06-041 blk 3.1 0.29 -25.53 0.18

-26.16PC06-041 blk rep1

1.1 0.29 -26.80 0.18

Mean -25.53 2.58 -25.60

Dri

llcor

e ke

roge

n

143.27m 2.2 0.27 -25.05 0.51-25.05143.27m rep1 3.5 0.31 -25.06 0.51

143.56m 10.6 0.50 -30.07 0.51-30.05143.56m rep1 13.2 0.68 -30.04 0.51

149.03m 3.6 0.24 -26.54 0.51-26.42149.03m rep1 1.9 0.22 -26.30 0.51

155.00m 9.2 0.26 -30.10 0.51-30.10155.00m rep1 8.9 0.29 -30.10 0.51

157.60m 18.9 15.52 -28.29 0.51

-28.03157.60m rep1 18.2 16.05 -28.30 0.51157.60m rep2 18.9 15.52 -27.80 0.51157.60m rep3 18.2 16.05 -27.73 0.51158.70m 21.0 12.41 -28.40 0.51

-28.38158.70m rep1 16.1 9.60 -28.14 0.51158.70m rep2 21.0 12.41 -28.28 0.51158.70m rep3 16.1 9.60 -28.70 0.51160.40m 15.8 14.36 -28.54 0.51

-28.29160.40m rep1 17.5 14.86 -28.74 0.51160.40m rep2 15.8 14.36 -27.98 0.51

160.40m rep3 17.5 14.86 -27.91 0.51160.70m 23.5 24.59 -28.22 0.51

-27.33160.70m rep1 23.5 24.59 -26.45 0.51162.40m 18.9 16.10 -27.80 0.51

-27.76

162.40m rep1 18.9 16.10 -27.72 0.51163.40m rep2 9.0 11.28 -27.62 0.51163.40m rep3 18.3 20.06 -27.83 0.51163.40m rep4 9.0 11.28 -28.45 0.51163.40m rep5 18.3 20.06 -27.13 0.51163.70m 17.2 15.77 -27.71 0.51

-27.77163.70m rep1 17.5 15.08 -27.72 0.51163.70m rep2 17.2 15.77 -27.77 0.51163.70m rep3 17.5 15.08 -27.88 0.51169.45m 9.1 0.61 -27.43 0.51

-28.36169.45m rep1 10.8 4.88 -28.23 0.51169.45m rep2 10.8 4.88 -29.41 0.51170.75m 12.0 10.06 -29.25 0.51

-28.59

170.75m rep1 11.2 1.58 -27.01 0.51170.75m rep2 12.4 1.69 -26.84 0.51170.75m rep3 12.0 10.06 -28.63 0.51170.75m rep4 11.2 1.58 -29.90 0.51170.75m rep5 12.4 1.69 -29.89 0.51288.55m 7.7 0.40 -30.08 0.51

-30.09288.55m rep1 5.5 0.27 -30.10 0.51310.53m 2.0 0.21 -30.11 0.51

-30.09310.53m rep1 3.4 0.53 -30.06 0.51311.64m 3.6 0.56 -13.70 0.51

-14.12311.64m rep1 3.7 0.54 -14.55 0.51312.32m 1.7 0.40 -17.56 0.51

-16.39312.32m rep1 2.8 0.52 -15.21 0.51313.37m 2.4 0.36 -18.72 0.51

-16.67313.37m rep1 6.5 0.49 -14.62 0.51Mean -26.78 4.30 -26.09

Met

apel

ite-h

oste

d ke

roge

n

PC06-024B 37.3 0.93 -17.22 0.20-16.21PC06-024B rep1 57.3 1.02 -15.21 0.20

PC06-024I 17.9 0.34 -21.88 0.20-22.03PC06-024I rep1 17.9 0.35 -22.17 0.20

PC06-027 4.7 0.32 -24.05 0.20-23.14PC06-027 rep1 20.4 0.32 -22.23 0.20

PC06-029 11.0 0.28 -25.74 0.20-25.58PC06-029 rep1 12.8 0.30 -25.42 0.20

PC06-031 8.3 0.62 -19.22 0.13-21.25PC06-031 rep1 1.6 0.35 -23.28 0.13

PC06-040 33.0 0.30 -24.58 0.13 -24.37PC06-040 rep1 42.1 0.35 -24.17 0.13

PC06-041 3.1 0.29 -25.53 0.13-26.16PC06-041 rep1 1.1 0.29 -26.80 0.13

PC06-044 3.1 0.27 -25.97 0.13-26.14PC06-044 rep1 3.2 0.27 -26.30 0.13

Mean -23.11 3.36 -23.11All Grand average -25.12 5.85 -24.66

Table 2. Coonterunah carbonate δ13Ccarb and δ18Ocarb data.Sample Size

(mg)CO2

yield (µmoles)

δ13Ccarb

(‰)± σ (‰)

δ18Ocarb

(‰, PDB)

± σ (‰)

Sample average (‰)

δ13Ccarb δ18Ocarb

93006 2.90 10.36 -1.94 0.13 -18.55 0.11-1.92 -18.6693006 rep1 3.20 12.80 -1.89 0.13 -18.76 0.15

94070 4.70 25.06 -3.03 0.12 -19.98 0.11 -3.03 -19.9894072 2.90 15.95 -3.89 0.13 -18.54 0.14 -3.89 -18.5494072 rep1 3.10 18.10 -3.88 0.08 -18.36 0.14 -3.89 -18.4596026B 15.00 75.20 -1.75 0.05 -7.38 0.13

-1.78 -7.2896026B rep1 13.10 62.83 -1.80 0.08 -7.18 0.0996033A 15.50 77.41 -3.27 0.08 -18.64 0.09

-3.27 -18.6496033B 15.80 84.71 -2.96 0.04 -13.71 0.0896033B rep1 20.40 20.40 -2.87 0.04 -12.82 0.09 -2.92 -13.2796045A 23.60 52.96 -0.88 0.10 -16.63 0.1

-0.89 -16.7796045A rep1 28.40 28.40 -0.89 0.05 -16.91 0.1396045B 10.30 48.27 -3.01 0.05 -13.81 0.09 -3.01 -13.81B55W 0.11 47.86 -3.12 0.03 -18.65 0.04 -3.12 -18.65B63W 0.19 70.01 -3.09 0.03 -17.99 0.03

-3.08 -17.94B63W rep1 0.19 64.47 -3.06 0.02 -17.89 0.04B75W 0.19 62.17 -3.54 0.03 -18.00 0.02 -3.54 -18.00B79W 0.73 86.94 -3.05 0.03 -17.41 0.02

-3.04 -17.40B79W rep1 0.54 84.79 -3.04 0.02 -17.39 0.03B80D 0.26 53.09 -3.30 0.04 -17.64 0.03

-3.25 -17.52B80D rep1 0.31 59.24 -3.20 0.03 -17.41 0.04B82W 0.30 72.48 -2.81 0.02 -16.73 0.02

-2.78 -16.55B82W rep1 0.17 49.24 -2.84 0.02 -16.60 0.03B82W rep2 1.27 86.63 -2.67 0.02 -16.30 0.02B87W 0.26 76.32 -2.66 0.02 -16.80 0.04

-2.69 -16.80B87W rep1 0.14 55.70 -2.71 0.02 -16.80 0.03B89W 0.17 36.93 -3.39 0.02 -18.36 0.04

-3.41 -18.38B89W rep1 0.12 23.39 -3.42 0.05 -18.41 0.05B90W 0.22 62.32 -2.85 0.01 -17.28 0.03

-2.85 -17.28B90W rep1 0.14 55.24 -2.86 0.02 -17.31 0.04B90W rep2 0.30 68.17 -2.84 0.02 -17.27 0.01PC05-020C 0.40 64.45 -3.12 0.02 -18.89 0.04

-3.13 -18.93PC05-020C rep1 0.58 69.99 -3.14 0.02 -18.98 0.03PC06-026 1.74 112.91 -3.29 0.04 -16.13 0.02

-3.24 -16.00PC06-026 rep1 1.01 106.77 -3.19 0.03 -15.86 0.02PC06-028A 0.67 42.09 -3.06 0.02 -19.45 0.02 -3.06 -19.45PC06-028B-i 0.71 74.74 -3.15 0.01 -19.35 0.04

-3.16 -19.36PC06-028B-i rep1

0.43 68.96 -3.16 0.02 -19.38 0.03

PC06-028B-ii 1.27 50.09 -3.04 0.01 -17.76 0.04 -3.03 -17.84PC06-028B-ii 1.28 47.12 -3.02 0.02 -17.92 0.04

450

rep1PC06-031 0.41 43.65 -2.49 0.04 -17.73 0.04

-2.48 -17.71PC06-031 rep1 2.00 48.00 -2.46 0.04 -17.68 0.03PC06-041 2.31 67.17 -3.36 0.03 -16.08 0.01

-3.34 -16.11PC06-041 rep1 1.69 67.52 -3.32 0.03 -16.13 0.03PC06-045 4.56 115.40 -1.96 0.01 -13.35 0.02

-2.03 -13.54PC06-045 rep1 2.72 121.12 -2.10 0.02 -13.72 0.01PC06-053 0.51 72.85 -2.70 0.02 -18.73 0.03

-2.70 -18.78PC06-053 rep1 0.30 70.28 -2.69 0.01 -18.84 0.03Mean -2.83 0.63 -16.91 2.61 -2.91 -17.10

Table 3. Coonterunah Δ13C data.Sample δ13Corg (‰) δ13Ccarb

(‰)Δ13C (‰)

96006K -26.65 -1.89 24.7696006K rep1 -23.1 -1.94 21.1696045AK -24.27 -0.88 23.3996045AK rep1 -22.87 -0.89 21.98

96045BK -28.77 -2.94 25.83PC05-020 avg -26.52 -3.13 23.39Mean (n=5) -25.36 -1.94 23.42Bulk (major mode)analyses

-25.12 -2.91 22.75

451452

453